Research map / Large language models / Retrieval and knowledge
ECG-QA: A Comprehensive Question Answering Dataset Combined With Electrocardiogram
Jungwoo Oh, Gyubok Lee, Seongsu Bae, Joon-myoung Kwon, Edward Choi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use pre-trained multimodal model (M3AE) instead of Fusion Transformer | ECG-QA baseline · ECG-QA test, S-Choose | EM accuracy: 46.4 -> 57.1 | Improved |
| use pre-trained multimodal model (MedViLL) instead of Fusion Transformer | ECG-QA baseline · ECG-QA test, S-Verify | EM accuracy: 72.1 -> 73.9 | Improved |
| remove ECG input (Blind Transformer) | Fusion Transformer · ECG-QA test, S-Verify | EM accuracy: 72.1 -> 67.7 | Hurt |
| remove question input (Deaf Transformer) | Fusion Transformer · ECG-QA test, CC-Verify | EM accuracy: 71.9 -> 59.5 | Hurt |
| use Verify instead of Query questions | M3AE · ECG-QA test | EM accuracy: 41.0 -> 75.5 | Improved |
| replace best QA model with best upper-bound model (SE-WRN) for perception | Single-Verify · ECG-QA Single-Verify | accuracy: 80.8 -> 83.1 | Improved |
| replace best QA model with MAX upper bound | Single-Verify · ECG-QA Single-Verify | AUROC: 0.808 -> 0.907 | Improved |
| replace trained QA model with zero-shot GPT-4 on SE-WRN text | ECG-QA · ECG-QA 10% test, S-Choose | EM accuracy: 58.2 -> 48.1 | Hurt |
| replace M3AE with SE-WRN + text-davinci-003 | ECG-QA · ECG-QA 10% test, S-Verify | EM accuracy: 76.0 -> 75.0 | Hurt |
| use gpt-4 instead of gpt-3.5-turbo | SE-WRN + LLM · ECG-QA 10% test, S-Choose | EM accuracy: 36.1 -> 48.1 | Improved |
| use gpt-4 vs majority baseline on CC-Verify | SE-WRN + gpt-4 · ECG-QA 10% test, CC-Verify | EM accuracy: 58.4 -> 54.9 | Hurt |
Closest papers by meaning
- anyECG-chat: A Generalist ECG-MLLM for Flexible ECG Input and Multi-Task Understanding (AAAI 2026 · similarity 0.88)
- ECG-Reasoning-Benchmark: A Benchmark for Evaluating Clinical Reasoning Capabilities in ECG Interpretation (NeurIPS 2026 · similarity 0.87)
- GEM: Empowering MLLM for Grounded ECG Understanding with Time Series and Images (NeurIPS 2025 · similarity 0.86)
- Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model (ICLR 2025 · similarity 0.86)
- ECG-R1: Protocol-Guided and Modality-Agnostic MLLM for Reliable ECG Interpretation (ICML 2026 · similarity 0.85)
- Zero-Shot ECG Classification with Multimodal Learning and Test-time Clinical Knowledge Enhancement (ICML 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
